A main steam line in P22 at 540°C is a solved problem. The same line at 570°C is a wall-thickness nightmare — and at 595°C, P22 is barely an option at all. Somewhere in that band, every power plant designer crosses to P91 and inherits a completely different set of problems in exchange.
The P22 vs P91 decision is really a trade: P91 buys you roughly double the creep strength and half the wall thickness, and charges you in fabrication discipline — tight heat-treatment windows, unforgiving welds, and a failure mode (Type IV cracking) that P22 barely knows. This guide maps where the crossover sits, what each grade costs in practice, the mistakes that have bent this decision’s history, and how to buy either grade safely.
The short answer: P22 (2¼Cr-1Mo) and P91 (9Cr-1Mo-V) are both ASTM A335 seamless alloy pipes for high-temperature service. The main difference between P22 and P91 is creep strength: P91’s vanadium- and niobium-strengthened martensitic structure retains roughly twice the allowable stress at 565–600°C, permitting much thinner walls — but it demands far stricter heat treatment and welding control. Below ~565°C, P22 is usually the economical choice; above it, P91 becomes progressively necessary.
What Separates P22 and P91 Metallurgically
A335 P22 carries 1.90–2.60% chromium and 0.87–1.13% molybdenum in a bainitic-ferritic structure, supplied normalized and tempered or annealed. It has been the workhorse of superheaters, reheaters, and hot steam piping for over sixty years — forgiving to weld, tolerant of field conditions, thoroughly understood. We compared it with its leaner sibling in our P11 vs P22 guide.
A335 P91 — “modified 9Cr-1Mo” — carries 8.0–9.5% chromium and ~1% molybdenum, plus the additions that change everything: vanadium (~0.20%), niobium (~0.08%), and controlled nitrogen. Normalizing around 1040–1080°C and tempering around 730–780°C produces a tempered martensite laced with fine vanadium-niobium carbonitrides. Those precipitates pin the microstructure against creep at temperatures where P22’s strength is collapsing.
That precipitate strengthening is the entire value proposition — and the entire vulnerability. Get the heat treatment right and P91 performs like a different class of material. Get it wrong, anywhere from mill to weld shop, and the strength you paid for silently isn’t there.
Why 565°C Is the Line
Plot allowable stress against temperature (ASME B31.1 / BPVC Section II-D) and the two curves tell the story. P22 holds respectable stress through the low 500s, then falls steeply. P91’s curve runs far higher and falls later. Around 565°C — the classic main-steam temperature of a modern subcritical/supercritical boundary plant — P91’s allowable stress is roughly double P22’s.
Double the allowable stress means, by the pressure-vessel arithmetic, roughly half the wall thickness for the same pressure and diameter. And wall thickness cascades: half the wall is half the weight per metre, smaller hangers and supports, and — critically for plants that cycle daily — much lower through-wall thermal gradients, which means less thermal fatigue every startup.
The P22 vs P91 Comparison Table
| Factor | A335 P22 | A335 P91 |
|---|---|---|
| Alloy system | 2¼Cr-1Mo, bainitic | 9Cr-1Mo-V-Nb-N, martensitic |
| Practical service sweet spot | Up to ~565°C | ~565–600°C+ |
| Relative allowable stress at 565°C | Baseline | ~2× |
| Wall thickness at high temp | Heavy | Roughly half of P22 |
| Weldability | Forgiving; standard preheat + PWHT | Demanding: strict preheat, tight PWHT window, hardness control |
| Signature damage risk | Temper embrittlement (long service) | Type IV cracking in weld HAZ |
| Field bending / repair tolerance | Good | Poor — deviations require full re-heat-treatment |
| Pipe cost per tonne | Lower | Higher — but fewer tonnes needed |
| Fabrication cost | Standard | Significantly higher |
| Industry track record | 60+ years, benign | Excellent when controlled; documented failures when not |
Where P91’s Reputation Was Earned — Both Ways
In practice, P91’s history carries a warning worth naming. During the grade’s rapid adoption from the 1990s onward, a wave of premature failures traced back to botched processing: pipe or fittings that missed the normalizing-and-tempering window, welds cooked above the lower transformation temperature during PWHT (creating fresh, untempered martensite), and soft-zone components with hardness far below specification. EPRI and industry bodies spent years documenting the lessons; hardness testing of P91 components became standard practice because of them.
The stable conclusion from that era: P91 is not a stronger P22 — it is a precision product. The specification is chemistry plus a heat-treatment recipe with narrow tolerances, and every downstream actor (mill, bender, welder, PWHT crew) can destroy what the melt shop created. A common check written into serious purchase specs today: hardness within roughly 190–250 HB, verified per component, because hardness is the cheapest window into whether the martensite was properly formed and tempered.
Type IV cracking deserves its own sentence. P91 welds develop a fine-grained heat-affected zone that creeps faster than parent metal; under long service, cracks form there preferentially. Designers manage it with weld placement, thickness margins, and inspection — but it means a P91 system’s life is governed by its welds in a way P22 systems largely aren’t.
The Decision Logic: Four Questions
- What is your design metal temperature? Below ~550°C: P22, almost without discussion — see our A106 vs A335 temperature-band logic for the ladder below that. Above ~580°C: P91 (or beyond — P92, at 600°C+ territory). Between: continue.
- Does your plant cycle? Daily two-shifting favors P91’s thin walls (lower thermal fatigue) — but demands honest weld inspection programs for Type IV. Base-load service weakens the case for the upgrade at the margin.
- Can your fabricator actually execute P91? Qualified procedures, calibrated PWHT with recorded time-temperature control, hardness verification, and no cowboy field bends. If the honest answer is no, heavy-wall P22 at 565°C is the safer engineering decision — a correctly built P22 system outlives a badly built P91 one.
- What does total installed cost say? P91 costs more per tonne but needs roughly half the tonnes at temperature, plus smaller supports; fabrication costs run the other way. At 565°C the totals often land close — which is why questions 2 and 3 usually decide.
Buying Either Grade: What Nakoda Steel Industry Puts in Writing
The verification stakes scale with the alloy. On P22 and P91 quotations, Nakoda Steel Industry commits to: dual A335/SA-335 certification with editions; full chemistry including the V-Nb-N window on P91 and residuals where temper-embrittlement limits (J-factor) apply on P22; furnace time-temperature charts for the normalize-and-temper cycle — for P91 these are not optional paperwork, they are the product; per-component hardness results on P91; and PMI at dispatch, because a mixed P22/P91 order is two identical-looking pipes with radically different futures. These are the same disciplines we detailed in our supplier red-flags guide — P91 simply raises the price of skipping them.
One honest scope note: P91’s fabrication risk lives mostly in your weld shop, not the mill. A supplier can hand you perfect pipe and a poor PWHT crew can still ruin the system. Buy the pipe verified; budget the fabrication discipline separately and seriously.
Common Mistakes in the P22-to-P91 Transition
- Specifying P91 for prestige at 540°C — paying precision-product prices where P22’s forgiveness is worth more than P91’s strength
- Treating P91 like “stronger P22” in the weld shop — the single root cause behind most documented failures
- Skipping hardness verification on delivered components and completed welds
- Field-bending or heating P91 without re-heat-treatment — any excursion near the transformation temperature resets the metallurgy
- Ignoring Type IV in life planning — inspecting parent metal while the welds govern system life
Frequently Asked Questions
What is the difference between P22 and P91 pipe?
Both are ASTM A335 seamless alloy pipes. P22 is 2¼Cr-1Mo bainitic steel, forgiving and proven to ~565°C. P91 is 9Cr-1Mo with vanadium, niobium and nitrogen forming a martensitic structure with roughly twice the creep strength at 565–600°C — at the cost of much stricter heat-treatment and welding requirements.
At what temperature should you switch from P22 to P91?
As a practical rule, P22 serves economically to about 565°C. Between roughly 550°C and 580°C the decision depends on cycling duty, fabricator capability, and wall-thickness economics; above ~580°C, P91 (and eventually P92) becomes the standard choice. Always confirm against current ASME allowable-stress tables for your exact conditions.
Why is P91 so difficult to weld?
P91 forms hard martensite on cooling and depends on a precise temper. Welding requires controlled preheat and interpass temperatures, and PWHT in a narrow window below the steel’s lower transformation temperature — overshooting creates fresh untempered martensite. Its welds also develop a fine-grained HAZ prone to Type IV creep cracking in long service.
Is P91 stronger than P22?
At elevated temperature, substantially — around double the allowable stress near 565°C, which permits roughly half the wall thickness. At low temperatures the difference matters much less, which is why P91 is rarely justified below ~550°C.
What is Type IV cracking in P91?
Creep cracking that forms preferentially in the fine-grained heat-affected zone beside P91 welds, where the microstructure is weakest under long-term stress at temperature. It makes weld placement, thickness margins, and periodic weld inspection central to P91 system life management.
What hardness should P91 pipe be?
Properly normalized-and-tempered P91 typically falls in roughly the 190–250 HB range. Values well below suggest a missed or excessive temper (soft, weak material); well above suggest untempered martensite. Serious purchase specifications require per-component hardness reporting — ask for it.
The Bottom Line on P22 vs P91
The P22 vs P91 decision is not “old grade vs better grade.” It is a forgiving material against a precision material, with 565°C as the practical hinge: below it, P22’s tolerance for real-world fabrication is usually worth more than P91’s strength; above it, physics stops giving you the choice. Whichever side of the line your project lands on, the pipe is only as good as its documented heat treatment — and that is a purchasing decision, made before the first weld.
Specifying main steam, superheater, or reheater piping? Send Nakoda Steel Industry your design conditions and line list — you’ll receive a P22/P91 quotation within 24 working hours with furnace records, hardness reporting, and PMI commitments stated in writing, and a straight answer about which side of 565°C your money should sit.